Vegetation: |
2.2 Monitoring methods 2.2.1 Survey design Climate change and drought literature provides few examples of field studies that
measure dieback of multi-layered sclerophyll vegetation communities analogous to the mallee and woodland communities on
Kangaroo Island. Viljoen (1995) designed field surveys to assess drought damage to selected woody species in Kruger National Park in South Africa. A belt transect method was used, with all individuals of selected species within a set distance of the transect line allocated to one of five canopy dieback classes. While this method allows flexibility in monitoring sites of varying size and habitat type, it is not suitable to detect dieback in understorey species, which was a principal focus in the Kangaroo Island study. The Kangaroo Island Vegetation Dieback Study targeted areas where vegetation dieback had previously been recorded (Mooney et al. 2007) but also included additional areas in order to monitor a range of vegetation communities across the landscape. To cover as much variability as resources would allow, sampling was stratified according to vegetation communities as defined by Ball and Carruthers (1998) and Regional Ecological Area (Table 3). Where possible sites were located in Department for Environment and Heritage (DEH) reserves, for ease of current and future access. Six sites surveyed were selected from dieback sites recorded in 2007, representing inland and coastal mallee and sheoak communities of the Eastern Plains and South Coast REAs (Be1, Wi1, De1, De2, Wc1, La1) (Table 3, Figure 4). The remaining two sites were selected to sample stringybark woodland and Eucalyptus remota communities of the Seddon and Gosse Plateau REAs (Pa1, Gl1). Sites selected to sample these latter two vegetation types, typical of the ironstone plateau landscape on Kangaroo Island, were located in Parndana Conservation Park and the eastern part of Flinders Chase National Park, known as the Gosselands. The Gosselands were burnt in a November 2002 bushfire, but were one of the few parts of the large reserve estate on western Kangaroo Island that was not burnt in the December 2007 bushfires. The 2007 fires were started by a series of dry lightning strikes and decimated around 40% of the remnant vegetation on Kangaroo Island. A macroplot design was favoured because larger plots allow better representative sampling of vegetation with a high degree of understorey variation and a sometimes scattered distribution of overstorey trees. Macroplots were randomly positioned in target vegetation communities within key areas, allowing inferences to be drawn about impacts across vegetation communities (after Underwood 1997). Macroplots are 100m x 50m in size. Plots were selected using GIS to generate numbered 100 x 50m cells in target vegetation community at each study site. Plots within 100 metres of the edge of a vegetation patch or a track were eliminated, except in the case of the two coastal plots at D Estrees Bay, where target vegetation was close to main D Estrees access track (De1, De2). The study plot was chosen by generating a random number. With the exception of the site on private property at Western Cove, the south-eastern corner of all macroplots is marked with a steel stake. Sampling took place over the spring and summer of 2008-09, commencing in November 2008, with the final two plots sampled in February 2009. 2.2.2 Understorey sampling: point intercept method The point intercept method used to sample ground cover, vegetation structure and floristics is an adaptation of the method developed in Kangaroo Island Biodiversity Monitoring Program (Pisanu et al. 2006). The key points of the field method are outlined below: (see Project Notes field for more details).
|